S3 with initial path

[petr tichý]

On S3, if user supplies the initial path, the request sent is

GET /<path>/?delimiter=%2F&max-keys=1000&prefix HTTP/1.1

that is GET Object: ​http://docs.aws.amazon.com/AmazonS3/latest/API/RESTObjectGET.html

but it should be

GET /?delimiter=%2F&max-keys=1000&prefix=<path> HTTP/1.1

GET Bucket: ​http://docs.aws.amazon.com/AmazonS3/latest/API/RESTBucketGET.html

to list the bucket starting with the prefix.

Also, the <path> should have the delimiter added at end.

[dkocher]

I cannot replicate this with the latest snapshot build. I have bookmarked the path /test.cyberduck.ch/test which leads to the request

GET /?delimiter=%2F&max-keys=1000&prefix=test%2F HTTP/1.1
Date: Sun, 21 Jul 2013 19:27:58 GMT
Host: test.cyberduck.ch.s3.amazonaws.com:443
Connection: Keep-Alive
User-Agent: Cyberduck/4.4 (Mac OS X/10.8.4) (x86_64)

[petr tichý]

Yes, but only when doing anonymous. When authentication is enabled, the requests are wrong. Tested on build 12380

[dkocher]

Fix and tests in r12437.